DrugBankNER

ETL of DrugBank to recognize KG2 concepts in DrugBank entries for use as training data

Prep work

1. Download DrugBank XML file from DrugBank.ca

   1. Make an account on DrugBank
   2. Run ./download_data.sh, which will put the DrugBank XML file in the data directory

2. You will need to have a copy of the RTX/ARAX node synonymizer

   1. Easiest way is to ask a team member for a copy
   2. Otherwise, you will need to still ask a team member to add your RSA key to the database server
   3. After that, you can get the sqlite file via

   scp [email protected]:/translator/data/orangeboard/databases/KG2.8.4/node_synonymizer_v1.0_KG2.8.4.sqlite .
   

   Note that this path is via [this line in config_dbs.json](https://github. com/RTXteam/RTX/blob/master/code/config_dbs.json#L3C28-L3C111) in case it gets updated

3. Set up the environment and install required packages:

   conda create --name drug_bank_NER python==3.11.10
   conda activate drug_bank_NER
   pip install xmltodict==0.14.2
   pip install pandas==2.2.3
   pip install spacy==3.8.2
   pip install scispacy==0.5.5

   Find your CUDA version by running:

   nvidia-smi

   Then, install the corresponding cupy-cuda package:

   pip install cupy-cuda<your_cuda_version>x

   Finally, download and install the ScispaCy models:

   pip install https://s3-us-west-2.amazonaws.com/ai2-s2-scispacy/releases/v0.5.3/en_core_sci_lg-0.5.3.tar.gz
   pip install https://s3-us-west-2.amazonaws.com/ai2-s2-scispacy/releases/v0.5.3/en_core_sci_scibert-0.5.3.tar.gz

Running the tool

1. Run perform_NER.py to perform named entity recognition and alignment on the text fields of DrugBank
2. Next, run look_for_identifiers.py to extract, synonymize, and align identifiers in DrugBank to RTX-KG2

The resulting NER and aligned results will then be in ./data/DrugBank_aligned_with_KG2.json